Svenja Granderath ; Prof. Dr. Jochen Dingfelder ; Dr. Peter Lewis

2019
Physikalisches Institut Bonn, Germany

Abstract: In this thesis an untagged exclusive analysis of the charmless semileptonic decay B→πlν using an integrated luminosity of 5.6 fb−1 of early Belle II data was discussed. The analysis was performed in 5 bins of momentum transfer, which makes it sensitive to form factor parametrisations. This method was chosen to prepare for later |Vub | extraction using a larger dataset. The signal selections were separately optimised in the momentum transfer bins using a figure of merit (FOM) based on the statistical significance of signal on top of background. First, selections based on detector properties and the kinematics of the signal decay were performed. After that, boosted decision trees (BDTs) were employed to separate signal from continuum background, as well as from BB background. After selecting on the BB BDT output classifier the shapes in signal and background of the signal extraction variables, M_bc and ∆E, were very similar and the correlations between the yields were high. Therefore, in the end, a selection was only placed on the continuum suppression BDT output classifier. Finally, we set up and validated a simultaneous extended binned maximum likelihood fit of the M_bc and ∆E distributions to extract signal in each momentum transfer bin. Fitting to data was not possible, due to very low statistics and high correlations in the yields of the templates. We performed toy studies to predict the luminosities necessary to achieve statistical significance of 3σ and 5σ in the total branching fraction for each of the modes. In the B0 modes a luminosity of ∼15 fb−1 and ∼20 fb−1 is needed for 3σ and 5σ significance, respectively. For the B± modes higher luminosities are needed: ∼30 fb−1 for 3σ significance and ∼50 fb−1 for 5σ significance

Note: Presented on 19 09 2019
Note: MSc